Blew controller by hooking battery up backwards

[fitek]

Hey folks, I ordered a bafang and two controllers from ecrazyman and promptly blew one up by attaching the batteries backwards. Stupid farking connectors... I chopped every single one off since finding matching connectors is impossible and of course while I was playing with the hall wires I didn't pay enough attention with the battery cables.

I took it apart and the main cap blew its top and the "r-sense" shunt has a gap.

What are my chances of resurrecting this controller or do I just order another one? I already replaced the cap and fixed the r-sense shunt but when i connect the battery I don't even get a spark of life. I'm assuming something else is blown but besides a few more caps it all looks solid state.

 

[The Stig]

Why cant they put one friggin diode in each controller to stop this from happening!!??

 

[theyerb]

Seriously, they need to put a diode in there. I did the same thing about 6 months ago. Luckily one of the traces on the board vaporized before any of the main components blew. I just had to connect the two ends of the trace with some solder, and voila- good as new. Have fun

 

[dnmun]

get real, why put a diode in the main current path?

you are really blaming the people who designed the board because he cannot tell positive from negative?

you really burned the shunt wires open?

we gotta get a homer simpson award.

i have the 100V1000uF caps, 30 cents each plus postage.

if the FETs are cooked, you could upgrade to the 100V 4110s, you could put some small caps back in temporarily and see if it will power up, were the phase wires and hall wires connected when you inverted it? if you can get 12V and 5V back, maybe the processor and the drivers survived. you can buy the shunts from digikey and mouser and newark to replace the melted ones.

so maybe you can build it back as a 100V controller , and can use the 72Vlifepo4 packs.

and pictures too! i have the 9FET infineon.

 

[icecube57]

I guess Ive been lucky. Ive done that on several occasions because i sometimes change my anderson configuration to gather data while charging through my watts up meter. I forget to swap the connectors back around. I get large spark and trip the BMS on battery. By the grace of the electrical gods I havent blown anything. I guess i have a diode in my controller.

 

[The Stig]

If there was a diode there would be no spark no current no nothing. Diodes only allow current to flow one way.
You have a BMS with short circuit protection and a very lucky controller!

 

[dnmun]

since you now are the designer, what diode would you install? what type of diode? what is the voltage drop across the diode while the controller is running, using 30A? how much power is lost across the diode? would it need a heat sink, and how large? when the flyback path sends motor back to the battery during regen, how does the current get to the battery?

its not as though the engineers don't think about his stuff, but how can you design something to prevent screw ups?

i still think it is possible the controller will run again. i suspect the processor does have protected inputs, not sure what would have happened to the voltage regulators.

fitek, power it up, look for the voltages on the 12V buss and the 5V buss. if there is not signal on the 12V buss, then try to use a 12V battery to apply a voltage to the output of the 12V regulator and see if the 5V regulator works by checking the voltage on the 5V buss. if there is 5V then maybe you could see if the processor works by cycling the hall signals with the wheel and then look for the controller to send out the signal to turn on the gate driver, those are ususally the first leads off the SOIC package along the top side adjacent the drivers.

if it is just the 12V regulator blown, then they are only 50 cents or so.

i really don't think the output FETs would blow, but they could i guess, maybe the high side flyback driver current caps could have blown too. do they show signs of venting? do the FETs look like they blew up? cracks or other obvious signs.

 

[The Stig]

I guess that answers my original question... They don't use diodes because they have resistance. I thought they had negligible resistance.

Those are some helpful diagnostics dnmun.

 

[kazbluesky]

Wait, I'm no electrician so forgive if I'm talking crap, but my 12 volt battery tender has reverse polarity protection - why not the same sort of thing for a controller?

 

[JinbaIttai]

Why engineer safety protection into something when we can instead rely on humans to do it perfect every time?

 

[huskydave]

If you install anderson connectors you won't have a problem in the future. It makes it very easy red to red black to black. You don't need diodes unless you are running 2 batteries of different chemistries or 2 batteries in series that have a bms. The diodes will protect your batteries under those conditions. Just be careful next time solder on some andersons and your good to go.

 

[fitek]

Why engineer safety protection into something when we can instead rely on humans to do it perfect every time?

The trouble is that the connectors on every controller are different and never match anything! If I hadn't been annoyed from farqing around with every other wire at the same time I would have paid more attention.

I'm surprised I haven't blown more controllers myself in the past few years. I have hooked up the batteries in other stupid configurations but never in reverse.

I'll take a look at the LM7805, I should have a few spares lying around. I hope this controller works but if it doesn't, at least I blew up one of the cheaper ones.

 

[fitek]

If you install anderson connectors you won't have a problem in the future. It makes it very easy red to red black to black. You don't need diodes unless you are running 2 batteries of different chemistries or 2 batteries in series that have a bms. The diodes will protect your batteries under those conditions. Just be careful next time solder on some andersons and your good to go.

I do have andersons, just I was moving too quick and put them on backwards. My brain registered what I was doing about a second before my hands completed the hook up.

 

[huskydave]

Understandable It is easy to make a mistake but if you learn from it (slow down) then you will have better luck in the future. I made many mistakes last year but the experience was worth all the headaches. Goodluck with the controller rebuild.

 

[The Stig]

Wait, I'm no electrician so forgive if I'm talking crap, but my 12 volt battery tender has reverse polarity protection - why not the same sort of thing for a controller?

Could anyone enlighten us as to what sort of reverse polarity protection kazbluesky's battery tender might have?

 

[dnmun]

his battery tender has a diode protecting the output, just like you asked should be included with the controller.

maybe i was a little rude, but the diode in the input has gotta been on their minds, and they don't use them.

but there is no way to do regen if you have diode in front of the battery, and i think it best to allow the battery to accept back emf from the flyback current to reduce the ringing of the voltage on the highside FETs. a diode would stop that current flowing back to the battery.

just my opinion of why there is no diode on the input, but i am just learning about controllers. but i have blown up things too, so didn't mean to end up self righteous.

i think it could be the miracle story if fitek can resurrect it from the dead, if the FETs did blow, which could very well be if it opened the shunt, but that should show up in a test, but if fitek can get the voltage regulators to produce 5V and if he has hall sensor voltage, the most critical thing would be if he could turn the drivers on and then if the FETs are bad, go with the 4110s or just some others. that was why i bot the 100V caps on ebay since i was gonna do the same thing with my 9FET infineon. methods did that and got tons of power out of them.

 

[fitek]

You can do it with a FET.

It's be a lot easier in lower power applications. Some of my guitar effects pedals work with DC brick supplies regardless of whether the power plug has + on the inner or outer part of the plug.

Google reverse polarity protection and you will get some circuits that work regardless of the polarity of the input.

This might not be a bad thing for us E-S folks to build. Certainly those of us who tinker around a lot could attach it to the controllers we want to protect.

 

[wrobinson0413]

.

 

[fitek]

Well I replaced the L7805CV with an LM7805 and the controller began behaving better. It didn't draw a bunch of current anyway. I got 5V on the + for the throttle and hall wires. But with all the wires attached, the motor doesn't even attempt to spin. Normally, even if the hall wires and phase wires all out of order, the motor will at least make noises, but nothing.

So I figure something is pretty dead in the controller. Unless anyone has any great ideas I'm going to leave this one for organ donation to some future controller.

 

[fitek]

Using mosfets in the return feed of the controller is the most efficient method for reverse voltage protection. The reason you don't see it on pretty much all chinese controllers is because the majority of people buy only the cheapest controllers. If you add the reverse voltage protection to a design, the controller is not going to be competitive with it's counter part that don't have this feature, and the company will lose money on the design. I think the Chinese try to mitigate the lack of RVP somewhat by putting polarized connectors on their controllers. So if you are making 5000 or 10000 ebikes for Walmart, the possibility that a person who works on the line will hook up a controller backwards with proper mating connectors is low. For someone who buys a controller on Ebay or such, and cuts the connectors off, the probability will be much higher that they will hook it up backwards, and be forced to buy more of the same product to replace it. Maybe that is the business model for selling into the DIY market :lol:

So the observation that can be made is that you get what you pay for. If you don't wish to pay for a good product, then you will have to accept the limitations of a cheaper lesser product.

This totally makes sense. What doesn't make sense is that someone buying it on ebay would probably pay the extra dollar or two to include the matching connectors. I mean, the motors and batteries I am using come with no connectors, so I use whatever I can scrounge. I've tried to find matching connectors through catalogs and by asking the folks at the Jameco counter but they all say it's pointless, just cut them all off and use one that match.

In fact, forget an extra $1 to $2. I would pay an extra $10 per controller to get matching connectors, but there's no such option.

 

[dnmun]

so the 12V regulator, the LM317T did not blow up? you only lost the 7805?

can you find any voltage on the highside rails to support the highside gate driver? i was thinking maybe the flyback caps sitting on the zener would have blown up but maybe the zener saved it. if you have 5V on the buss, i would recommend looking for a gate driver signal from the processor when the wheel is moved to cause the hall sensors to signal the processor.

even without the FETs working, it is possible to check to see if the driver circuits work and then you could figure out if the FETs blew up too.

the little 6FET infineon will put out 2.4 kW@72V lifepo4 when the shunt is soldered up and the FETs go to 4110s, and 100V1000uF caps for the input and S/D busses.

you could end up with a superior controller from what you now consider useless trash. check your gate drivers for signals.

 

[fitek]

I checked the 12v regulator. It's not putting out 12v because the input is only 9v. Something like 8.6 volts actually makes it to the LM7805, which does put out 5v (not surprising since I changed it).

The only cap that looked damaged was the main cap, since replaced. The top blew open and there was fuzzy stringy stuff hanging out. I didn't have the same size on hand but something fairly close; good enough for testing.

Actually when I attach the battery, the big resistor R1 that feeds the logic side heats up, but slowly. Last night I hadn't left it connected long enough to notice. So it seems like something is shorted on the logic side.

I measured the resistance of the FETs in circuit. If the legs are numbered 1 2 3 going from left to right when looking at the label on the fets, 1 and 2 seem to be in meg ohms and 2 and 3 in kohms except for the #5 fet (again, counting left to right) which is 3k ohms across both legs.

 

[fitek]

I bought another controller and was up and running in 30 seconds. The performance with the new controller was pretty cool.

This morning, the bike screwed me three separate times though. First, throttle trouble, so I had to go back home and fix it. In a rush to catch the next train, it threw the chain and it got stuck against the frame.

Finally, the motor died a block later. There is a lot of resistance when turning it, and applying power has no effect. If it hadn't been for the $50 controller inside it, I would have just left it there by the side of the road since I was already late to work and really needed to catch the next train. So I had to put gas in my car for the first time in almost two months and just drive. Originally left the house at 7:30am and its almost 10am now.

I guess that's what you get when you trust a cheap Chinese ebike. I've always found a way to at least limp my ebikes home.

 

[dnmun]

sorry about your troubles. i never figured why your 12V regulator was only seeing 9V input originally. but you should have been able to see the 5V on the hall sensors and the 12V for the driver circuitry too. now this controller is dead or just wires disconnected?

 

[philf]

The proper place for a diode in this scenario is not inline with the flow of power, but rather ACROSS it. When things are hooked up correctly, the diode is transparent and nothing happens. When reverse polarity is applied, the diode conducts and pops the fuse. THAT is the kind of circuit protection that Kaz alludes to having in his charger...